ORTHOPAEDICS e II: SPINE AND PELVIS
Discitis and spinal
infection
Edward Matthews
Oliver Stokes
Abstract
Spinal infection poses a diagnostic challenge and a low threshold for
investigation should be maintained. Presentation is varied and non-
specific symptoms mean that patients are investigated by many spe-
cialities. The majority of spinal infection is from haematogenous
spread and therefore an origin of infection needs to be sought. Treat-
ment of spondylodiscitis is routinely managed by non-surgical treat-
ment with a prolonged period of antibiotics. Complications of
spondylodiscitis can lead to morbidity and may be difficult to treat
and often require surgery. It is essential to attempt to obtain microbi-
ological diagnosis. Initial management and investigation does affect
treatment strategies and it is important to understand this.
Keywords Discitis; epidural abscess; osteomyelitis; spinal infection;
spondylodiscitis; vertebral
Introduction
Adult spinal infection is an uncommon clinical condition and has
the potential to cause a spectrum of morbidity and occasionally
mortality. This article provides the reader with an understanding
of spinal infection and an update on current guidelines and
evidence-based best practice.
Spinal infections can range from indolent to rapidly indolent
or rapidly destructive. Immunocompromised patients are
particularly at risk, and spinal infections can present to a wide
range of surgical and medical specialities. The natural history of
the disease is dependent on host, pathogen and comorbid vari-
ables and can lead to progressive destruction which if untreated
can result in significant deformity, neural compromise and death.
The clinical picture on presentation or diagnosis affects subse-
quent investigation and management. Therefore, thorough
assessment of these patients is essential. Spinal infections most
often occur due to haematogenous spread of infection from
elsewhere in the body; de novo spinal infection is rare, therefore
multisystem assessment of these patients is important.
There are a number of descriptive medical names that are
synonymous for the same clinical entity. Discitis refers to an
infective condition of the intervertebral disc. Commonly this is
used to describe pyogenic spondylodiscitis, taking the prefix from
the Greek, spondylos, for vertebra, a clinical condition that
Edward Matthews FRCS(Tr & Orth) is a Trauma and Orthopaedic
Specialist Registrar in the South West Deanery, UK. Conflicts of
interest: none.
Oliver Stokes FRCS (Tr & Orth) MSc is a Consultant Spinal Surgeon at
the Royal Devon and Exeter Hospital, Exeter, UK. Conflicts of
interest: none.
SURGERY 36:7
involves a pus-producing infection of the disc and vertebrae. It can
be considered that spondylitis, discitis and spondylodiscitis are a
continuum of the same process.1 This clinical spectrum can also
involve primary or more commonly secondary epidural abscess,
pyogenic facet joint septic arthritis and vertebral osteomyelitis.
Classification and epidemiology
Spinal infection can be characterized by the immune reaction to
the causative organism. In the developed world the majority
cause a pyogenic reaction. The developing world has a higher
burden of disease from Mycobacterium tuberculosis (TB) and
zoonotic infections.
Meticillin sensitive Staphylococcus aureus (MSSA) is the
most common organism isolated (63%) and along with
Streptococcus species (20%)1 cause the majority of infections.
However, in 25e33% of patients no pathogen is isolated.1,2 Other
typical organisms include Gram negative bacilli, commonly seen
in infections secondary to intravenous drug abuse. Salmonella
infection is more common in those patients with immune
compromise and sickle cell anaemia.2 Table 1 provides a list of
commonly encountered pathogens and their prevalence.
Granulomatous discitis occurs most commonly from M.
tuberculosis (TB) and brucellosis. These combined with Candida
spp. and other fungal infections cause significant morbidity and
mortality in the immunosuppressed and in the developing world.
Brucella infection, a zoonosis, occurs secondary to consuming
unpasteurized dairy products or occupational exposure to
infected animals.
Spondylodiscitis accounts for a small burden of disease in the
developed healthcare system and has been reported to represent
2e7% of all cases of osteomyelitis.2 The burden of spondylo-
discitis is reported to have an incidence of 3.7/100,000 per year
according to recent UK-based study.3
Factors that increase this risk of developing discitis in the UK
are widely reported to be diabetes mellitus, steroid use and
immunosuppression, intravenous drug abuse, malnutrition and
renal failure. These are summarized in Box 1. Tuberculosis and
human immunodeficiency virus (HIV) pose a significant burden
of disease in the developing world. Pyogenic spondylitis is more
common in the elderly, whereas epidural abscess is more of a
problem with disseminated bacteraemia and iatrogenic causes.
Hadjipavlou et al.1 found septicaemia to be the biggest risk factor
for spondylodiscitis, with tobacco and intravenous drug use
being the next most common risk factor (Table 1 and Box 1).
Pathogenesis
The aetiology of spinal infection is either haematogenous spread,
iatrogenic or contiguous. Haematogenous infection can be
disseminated by either the arterial tree or the venous system.
This is the most commonly encountered source in clinical prac-
tice. The arterial vascular supply to the spine is via segmental
arteries. The intervertebral disc itself is avascular (in adults). The
vertebral endplate, however, is vascularized, allowing for direct
extension from endplate to disc, hence the surrounding end
plates are involved in cases of discitis, leading to the term
spondylodiscitis the term spondylodiscitis is often used. Prostatic
biopsy can lead to a venous source of infection via the valveless
vertebral veins of Batson. Infective endocarditis and urinary tract
352 Ó 2018 Published by Elsevier Ltd.
 ORTHOPAEDICS e II: SPINE AND PELVIS

They also recommended that neurological symptoms with

Pathogens1,4e7 fever should prompt urgent investigation.8
Organisms Percentages Risk factor Unremitting pain is a significant feature and is present in 90% of
presentations.2 Red flag symptoms should specifically be elicited
Staphylococcus sp. 63e20 and if present result in expedited investigation and treatment. Box 2
MSSA 36e21 Surgery is a commonly described red flag symptoms that should be spe-
MRSA 14e7 Elderly cifically asked for in the history; it is not exhaustive but does
Coagulase negative 3e16 Device related prompt suspicion for consideration of serious diagnoses. The
staphylococci presence of neurological deficit is a bad prognostic sign and implies
Streptococcus spp 19e6 a compressive or vascular lesion and should arouse suspicion of an
Escherichia coli 11e4 UTI epidural abscess, it should be treated as a surgical emergency.
Pseudomonas aeruginosa 6e3 IVDU Neurological deficit is present in one-third of patients with an
Salmonella spp. 2 Sickle cell epidural abscess.2 Epidural abscesses, either primary or sec-
Propionibacterium sp. 1 Postoperative ondary from spondylodiscitis, are reported to be present in 35%
Brucellosis 21e48 Developing world e of all spinal pyogenic infections.1 Severe neurological deficits are
unpasteurized goats milk present in 37%, including paraplegia or paralysis; with surgical
Tuberculosis 9e46 Developing world decompression only 23% of those make complete recovery.2
No organism cultured 24 Postoperative infection has a quoted incidence of 2% following
lumbar discectomy.9 The most common presentation of these
IVDU, intravenous drug use; MRSA, meticillin sensitive Staphylococcus aureus;
MSSA, meticillin sensitive Staphylococcus aureus; UTI, urinary tract infection. cases are acute postoperative surgical site infections, with the
characteristic signs of acute inflammation, wound discharge and
Table 1 raised inflammatory markers. These cases are a distinct group
from other causes of discitis and should be managed aggressively
with a low threshold for debridement and intervention. Further
detail on postoperative discitis is beyond the scope of this article.
Risk factors for discitis
Examine for the presence of deformity, para-spinal spasm and
referred pain. Examination findings should demonstrate or
C Intravenous drug abuse
exclude fever and systemic sepsis. If a large collection is also
C Elderly
present, then psoas and a para-spinal abscess may co-exist.
C Chronic renal failure including renal replacement therapy
Document the extent of neurological deficit. The examination
C Infective endocarditis
should be repeated at regular intervals. Progressive neurological
C Diabetes mellitus
deterioration requires emergency surgical intervention. The
C Steroids & immunosuppression
American Spinal Injury Association impairment scale is a useful
C Malnutrition
tool for confirming the neurological deficit and charting any
C Tuberculosis and human immunodeficiency virus
progression.10
Box 1
Investigation
infection have been reported to be present in 12e17% of spon- Initial investigation guides early management of disease process.
dylodiscitis diagnoses, respectively.7 The recommended initial diagnostic tests include the presence of
Iatrogenic infection complicates 1e15% of lumbar spine sur- raised inflammatory markers and positive results of magnetic
geries. Risk factors include diabetes, revision surgery and post- resonance scanning of the whole spine.
operative haematoma. Epidural anaesthesia and lumbar punc- Testing for raised inflammatory markers has a low cost and
ture also expose patients to the risk of discitis, as do spinal cord has been shown to have a sensitivity of 94e100%.8 Obtaining
stimulators.
Direct or contiguous spread is rare. It can occur with cases of
mycotic aneurysm, oesophageal pathology, retropharyngeal ab- Red flag signs
scess and intraabdominal sepsis.
C Weight loss
Presentation C Thoracic back pain
C Sphincter dysfunction
Diagnosis of spondylodiscitis requires a high index of suspicion. C Neurological deficit
Delay in treatment may cause significant complications and C Age <20 or >55 years
morbidity. Recent guidance5 from the Infectious Diseases Society C Known history malignancy or trauma
of America (IDSA) suggests that the diagnosis should be sus- C Night pain or non-mechanical pain
pected in those with: C Fever
 new or worsening back pain C Corticosteroid use or immunosuppression
 back pain associated with fever or raised inflammatory C Structural deformity
markers or recent staph aureus infection/infective endo-
carditis/septicaemia. Box 2

SURGERY 36:7 353 Ó 2018 Published by Elsevier Ltd.

 ORTHOPAEDICS e II: SPINE AND PELVIS

a culture of the causative organism guides effective treatment

and influences prognosis. It is for this reason that the two sets of
bacterial blood cultures (both aerobic and anaerobic) are rec-
ommended when the diagnosis of spondylodiscitis is
considered.8
It is commonly accepted that prior to starting empiric antibi-
otics, the responsible organism is cultured. Blood culture is an
inexpensive and effective method of obtaining this. If this fails to
yield a culture, then a percutaneous biopsy is recommended.
However, when the patient has a significant infection the ability
to safely delay treatment to obtain a microbiological diagnosis
should be abandoned if there is evidence of:
 systemic sepsis
 haemodynamic instability
 neurological compromise/cauda equina syndrome (where
immediate surgical decompression is undertaken).
The use of antibiotics reduces the microbiological yield of
biopsy and blood culture and for this reason and in the absence
of the clinical scenarios above it is common practise to withhold
antibiotic use until a microbiological sample is obtained.
Magnetic resonance imaging (MRI) is the radiological investi-
gation of choice to help make the diagnosis. MRI is widely avail-
able during working hours and has a sensitivity of 97%, specificity
of 93%, and an accuracy of 94% in diagnosing spondylodiscitis.8
MRI can also differentiate between infective and neoplastic pro-
cesses. Where negative, and the diagnosis of discitis is still
strongly suspected, a serial or follow up scan is recommended at a
3-week intervals. Figure 1 shows the T2 mid sagittal sequence of
spondylodiscitis at L4/5 which demonstrates the hyper-intense
signal in T2 and also demonstrates epidural abscess posteriorly.
Figure 2 shows the T1 mid sagittal sequence of multilevel thoracic
discitis with fluid collection anteriorly to this.
 The specific findings on MRI in spondylodiscitis is the high Figure 1 T2 mid-sagittal sequence of spondylodiscitis at L4/5.
signal on T2-weighted images and confluence of disc space
and vertebral marrow on T1 images
 Sub-ligamentous spread and the involvement of more than
three vertebral levels is suggestive of tuberculous
infection11
The use of contrast agents can be helpful in some circum-
stances. Gadolinium-based contrast MR imaging can be useful in
the differentiation of infected granulation tissue from abscess.
The contrast will show a peri-abscess halo of intense signal
surrounding an area of low signal that will not be present in
granulation tissue. The use of gadolinium, as with all contrast,
comes with the risk of nephritis and therefore should be used
with appropriate caution after considering the risks and benefits.
Computed tomography has a significant role to play where
biopsy is required. Image-guided aspiration is recommended
where the causative organism has not been confirmed by either
blood cultures or serological testing.8 It is advised against when
there is a positive bacteraemia involving S. aureus, Staphylo-
coccus lugdenesis or Brucella. Biopsy results should be sent for
microbiological culture and pathological review to exclude
neoplastic process.
Bone scanning can be combined with gallium scans to give a
sensitive investigation of >90% for discitis where MRI cannot be Figure 2 T1 mid-sagittal sequence of multilevel thoracic discitis with
employed. White-cell labelled scanning is not recommend.8 fluid collection.

SURGERY 36:7 354 Ó 2018 Published by Elsevier Ltd.

 ORTHOPAEDICS e II: SPINE AND PELVIS
Plain radiographs are not useful in the acute diagnostic work
up; it is widely accepted that there is a period of 4 weeks in which
radiographic signs will not be present in spondylodiscitis, if at
all. The common findings are a destructive process that is centred
around the disc space, i.e. the superior endplate and the inferior
end of the vertebral body above. Due to a destructive appear-
ance, it is also important to rule out a neoplastic process and a
myeloma screen including serum electrophoresis and urinary
Bence-Jones proteins to diagnose myeloma should be performed.
It is important to consider the aetiology of discitis when a
diagnosis is established. Careful evaluation of risk factors could
point to infective endocarditis, intravenous catheter infection,
solid organ abscess and haemodialysis, all of which have sig-
nificant morbidity and mortality if not recognized and treated.
Management
The management can be complex and needs tailoring to the
patient’s clinical picture.
Antibiotics
The mainstay of treatment is a course of intravenous antibiotics
until inflammatory markers are on a downward trend and
heading towards a C-reactive protein (CRP) below 50 mg/l.
CRP of less than 30 mg/l is common around the 4-week point,
following treatment, according to Sur et al.3 There is also no
consensus on the length of intravenous therapy and on the
conversion to oral therapy. There is evidence that 6 weeks of
antibiotic therapy has no inferiority to 3 months of antibiotics.12
The current consensus on empirical antibiotic therapy is that
there should be careful evaluation of patient background, past
medical history and social and geographic susceptibility.8 Where
fungal or mycobacterial pathogens are suspected there is a
limited role for empirical therapy.
The main pyogenic organisms are MSSA, meticillin resistant
S. aureus (MRSA) and streptococcal species and Gram negative
species; local empiric guidance will involve broad-spectrum an-
tibiotics until culture positive guidance agent can be employed.
Surgical intervention
There is limited evidence-based guidance to guide surgical
management of spinal infection. The indications emergency
decompression and washout include:
 systemic toxicity from sepsis
 progressive neurological compromise (including cauda
equina syndrome and cord compression).
Indications for less emergent surgical intervention include:
 medical comorbidity with inadequate response to medical
therapy - percutaneous washout with Jamshidi needle
 vertebral collapse
 unremitting back pain and nerve root/foraminal
compression, as sequelae of discitis
 progressive deformity.
The role of surgery is reserved for when medical therapy is
not appropriate (detailed above) or when it has failed. The risk
factors for failed medical management of epidural abscesses
have been reported by Kim et al.: patients older than 65 years
of age, MRSA infection, diabetes mellitus and neurological
symptoms.13
SURGERY 36:7
Bracing
Orthotic use in spinal infection has a limited evidence base.
Where the presence of infection leads to deformity and spinal
instability the use of an orthotic is likely to be insufficient apart
from a temporary device to give comfort, and not as a long-term
solution. Surgery for kyphosis and pseudarthrosis provides a
better outcomes and lower pain scores.1
Surveillance
Patients should be reviewed regularly to evaluate clinical
improvement in pain and systemic symptoms. Clinical exami-
nation should also be regularly repeated to assess for neurolog-
ical sequelae and development of deformity. The use of serial
inflammatory markers is widely reported but there is limited
evidence on to how to interpret slowly declining CRP. It has been
reported that a CRP of 27.5 mg/l at 4 weeks could be a predictive
marker of how effective treatment is but there is a paucity of
evidence for this.6
Outcomes
If patients suffer from a complication of discitis they have a poorer
outcome. The potential for paralysis and deformity pose signifi-
cant morbidity. Absolute failure can cause, at best, deformity and
pain with ongoing osteomyelitis and, at worst, it can cause death
from sepsis. Figure 2 demonstrates an inadequately treated spinal
infection in a patient showing significant destruction and
deformity.
Epidural abscess formation secondary to discitis is a common
complication of infection, affecting 35% in one series,1 and one-
third these cases were associated with paralysis.
Mortality from discitis is rare. A recent case series from
Madrid reported a mortality of 7/108 (6.5%).14 The 1-year
mortality has been reported to be 11%,15 but attributing this
mortality rate solely to discitis is difficult due to the underlying
comorbidities present in many affected patients. Comorbid pa-
tients represent the more likely to be affected and they are likely
to have the worst outcomes.
Conclusion
Spinal infection is a condition that presents to many specialities
and in many clinical entities, without pathognomonic clinical
symptoms or signs. Detection requires a high index of suspicion.
Initial management has significant implications as to subsequent
treatment. Poorly treated spondylodiscitis can have serious
consequences and morbidity and therefore is best performed by a
multidisciplinary team of physicians, surgeons, radiologists and
microbiologists. A
REFERENCES
1 Hadjipavlou AG, Mader JT, Necessary JT, Muffoletto AJ. Hema-
togenous pyogenic spinal infections and their surgical manage-
ment. Spine 2000 Jul 1; 25: 1668e79.
2 Govender S. Spinal infections. J Bone Jt Surg Br Vol 2005 Nov 1;
87-B: 1454e8.
3 Sur A, Tsang K, Brown M, Tzerakis N. Management of adult
spontaneous spondylodiscitis and its rising incidence. Ann R Coll
Surg Engl 2015 Sep; 97: 451e5.
355 Ó 2018 Published by Elsevier Ltd.
 ORTHOPAEDICS e II: SPINE AND PELVIS
4 Skaf GS, Domloj NT, Fehlings MG, et al. Pyogenic spondylo-
discitis: an overview. J Infect Public Health 2010; 3: 5e16. King
Saud Bin Abdulaziz University for Health Sciences.
5 Bhavan KP, Marschall J, Olsen MA, Fraser VJ, Wright NM,
Warren DK. The epidemiology of hematogenous vertebral osteo-
myelitis: a cohort study in a tertiary care hospital. BMC Infect Dis
2010 Jun 7; 10: 198. BioMed Central.
6 Yoon SH, Chung SK, Kim K-J, Kim H-J, Jin YJ, Kim HB. Pyogenic
vertebral osteomyelitis: identification of microorganism and lab-
oratory markers used to predict clinical outcome. Eur Spine J
2009 Nov 24; 19: 575e82.
7 Gouliouris T, Aliyu SH, Brown NM. Spondylodiscitis: update on
diagnosis and management. J Antimicrob Chemother 2010 Sep
28; 65(suppl 3): iii11e24.
8 Berbari EF, Kanj SS, Kowalski TJ, et al. 2015 Infectious Diseases
Society of America (IDSA) clinical practice guidelines for the
diagnosis and treatment of native vertebral osteomyelitis in adults,
vol. 61. Clinical Infectious Diseases : an Official Publication of the
Infectious Diseases Society of America, 2015; e26e46.
9 Weinstein JN, Tosteson TD, Lurie JD, et al. Surgical vs nonoper-
ative treatment for lumbar disk herniation: the Spine Patient
Outcomes Research Trial (SPORT): a randomized trial. JAMA,
J Am Med Assoc 2006 Nov 22; 296: 2441e50.
SURGERY 36:7
10 Jha AASIA. Impairment scale. In: Encyclopedia of clinical neuro-
psychology. New York, NY: Springer, 2011; 255e7.
11 Hong SH, Choi J-Y, Lee JW, Kim NR, Choi J-A, Kang HS. MR
imaging assessment of the spine: infection or an imitation? Ra-
dioGraphics. Radiol Soc North Am 2009 Mar; 29: 599e612.
12 Bernard L, Dinh A, Ghout I, et al. Antibiotic treatment for 6 weeks
versus 12 weeks in patients with pyogenic vertebral osteomyelitis:
an open-label, non-inferiority, randomised, controlled trial. Lancet
2015; 385: 875e82.
13 Do Kim S, Melikian R, Ju KL, et al. Independent predictors of
failure of nonoperative management of spinal epidural abscesses.
Spine J 2014 Aug 1; 14: 1673e9. Elsevier Inc.
14 Cebria
n Parra JL, Saez-Arenillas Martín A, Urda Martínez-
Aedo AL, Soler Ivan~ez I, Agreda E, Lopez-Duran Stern L. Man-
agement of infectious discitis. Outcome in one hundred and eight
patients in a University Hospital. Int Orthop (SICOT) 2012 Jan 4;
36: 239e44. Springer-Verlag.
15 Kehrer M, Pedersen C, Gorm Jensen T, Hallas J, Lassen AT.
Increased short- and long-term mortality among patients with
infectious spondylodiscitis compared with a reference population.
Spine J 2015 Jun 1; 15: 1233e40. Elsevier Inc.
356 Ó 2018 Published by Elsevier Ltd.